ARCHITECTURE IS THE KEY TO UNLOCKING THE BENEFITS OF IP VIDEO

by Barton Kartoz

In the late-1980’s, Intel and Microsoft all but destroyed Digital Equipment Corporation (DEC) by enabling desktops that could distribute computing and storage power to each user. DEC continued to assume that computer efficiency would be maximized through centralization. It appears that a similar battle is being waged today about the architecture needed to promote effective Internet Protocol (IP) video solutions today. Will centralized architectures, which mirror what was possible in the past, continue to be the de-facto standard into the 21st century? Or, will a new architecture of highly distributed, “smarter” IP devices enable a future of increasing video quality and performance?

First, what are the real value propositions of IP video Why has the US market been slower to adopt than many anticipated?

The most significant value propositions of IP video include:

1. Standardized, shared pipe - Instead of isolated signals between each camera and a recorder, IP allows multiple signals to travel on a single cable or channel. This has significant implications for ease of deployment, installation cost, and long term maintenance of the transmission/cable system.

2. Shared Resources – Almost all organizations need an IP infrastructure for computing, communication, and business data. Video can be one more use for this infrastructure, rather than being completely separate.

3. Quality – The NTSC standards have constrained analog video quality to 4 CIF, (708 x 480 @ 30 feet per second (fps)) or lower since their introduction in the middle of the last century. IP standards do not limit the image size or frame rate as analog systems do. This translates into options for higher quality.

4. “Smarter” – As cameras become IP devices, we gain potential to add additional computing tasks in addition to capture and transmission of video. Examples: object and character recognition and advanced motion analysis. This “intelligence” has the promise to make video more useful.

Misinformation still abounds. One major video software developer’s web site suggests that when video quality is a primary concern, clients should consider an analog system. Why? Because implementation of IP video to date has often resulted in lower video quality! This perspective, legitimately based on early experience with poor IP video picture quality and management, can limit anticipated acceptance rate for IP video. If IP has the promise of higher quality, more function, and lower long term cost, but real outcomes are poorer quality, then who would blame users for balking?

I believe it’s largely an architectural issue. Historically, video systems have always been a centralized command and control product. From the advent of video security in the 1930’s through the 1990’s, video required home-run cabling to centralized locations.

A large majority of the video products industry is still taking a centralized approach today. Beefy servers are being deployed with large numbers of cameras streaming inbound for simultaneous storage. Video analysis tools are being placed on secondary, also beefy servers. While this may be a natural and understandable evolution from the analog architecture(s) of the past, quality is being sacrificed for perceived efficiency, and conformance to established architecture. Certainly efficiency is important, but are centralized systems truly more efficient?

Most other network-based systems that we encounter in our daily lives are decentralized. Consider your desktop or laptop, and the network it connects to. Another example close to
home is the access control industry. Access control system makers long ago realized that a distributed architecture would promote fault tolerance, higher performance, and more features. Video solutions are one of the most use-intensive services on our networks. Centralized approaches lead to even more reliance on the network. Viewed in this light, it seems surprising that video systems would ever be deployed in highly centralized designs.

All of this would be understandable if centralization were the only cost effective way to deploy. This is not the case. Technology enables a sweeping change from centralized to distributed systems. By distributing the computing, storage and analysis associated with video, all of the IP value propositions discussed above can be fully realized.

Some key attributes of a distributed system include:

1. Moving decision-making analytics away from centralized servers and onto the peripheral sensors or cameras they serve. It turns out that on top of de-coupling network uptime and throughput from video system reliability, this change generally lowers cost as well.

2. Keep streaming video off networks except for when it is being used.

3. Record at the very highest possible quality without impact to the network. Stream live view video at lower quality levels in general, but reserve enough network capacity to stream highest quality on demand.

4. Use of imbedded operating systems that are less prone to network based threats.

Distributed systems have other benefits. Properly engineered, these will also apply to video implementations:

1. Reduce impact of equipment failures. A single failure should only impact a small section of a distributed system.

2. Increase overall uptime of the system by removing network and centralized services as causes for recording and alarm generation failures. With the performance and cost effectiveness that is becoming available utilizing IP technology, there is little doubt that analog video will not be the dominant system in the future. How far into the future depends on when IP solutions rise above their predecessors in terms of usability and performance. This largely depends upon our collective willingness to embrace a major change in the way we design and implement video systems.

Barton Kartoz is a Regional Manager for CoVi Technologies, a market leader in distributed, high definition (HD) IP video solutions. Kartoz, a US Navy veteran, did his BA work at University of Pennsylvania in Biochemistry and holds an MBA from Rutgers University. He is an ASIS Central NJ Chapter member and can be reached at bkartoz@covitechnologies.com

>> Return to This Issue's Table of Contents